environment affect true intelligence?" or "what proportion of institutionalized persons are pseudo mental defectives?") than have been devoted to the reporting of research investigations.
The third construction for both language and mental deficiency leads to a different set of questions. With regard to language the first question might be, "To what behaviors shall the term language refer?" This question is to be answered arbitrarily; nevertheless, the individual examiner can answer it for the purposes of his own research. The second question might be, "In what way are the classes of behaviors included under the label 'language' related to each other?" Finally, "What are the variables which control the various classes of language behavior?" Specific research questions based on the above formulation would lead directly to a body of knowledge concerning the prediction and control of language behavior based on a systematic program of manipulation of the relevant variables.
If mental deficiency were similarly construed, a similar body of knowledge concerning prediction and control would be obtained.
The approach just presented is not new. Skinner ( 1957), Kantor ( 1952), and Turner ( 1961) have all made statements similar to those made by the present author. This fact, of course, does not indicate endorsement by Skinner, Kantor, and Turner of statements made in this chapter.
Within this chapter language referred to speech and gestures and to the reactions of another person to speech and gestures. A review of research on speech and gestures, the listener's understanding of speech and gestures, and the interaction of speaker and listener followed from this conception of language.
A review of new language evaluation procedures revealed that most of the procedures included provisions for sampling both speech and speech comprehension. In addition, two tests (the PLS and the ITPA) yielded provisions for sampling either gestures or the comprehension of gestures. The PLS and ITPA are both based on learning models and hold promise not only for sampling language in a better organized fashion but also for bringing the study of language in line with broader general behavior systems. The tests present two possible limitations. First, although both tests are based on learning models, neither provides any measure of current rate of acquisition. Second, both obtain their samples of language in a formal situation with an adult examiner. These samples may or may not be representative of language samples obtained in other situations.
There have been numerous studies of the prevalence of speech defects among mentally defective populations. In spite of the use of differing populations of defectives and differing criteria for speech deficits, the results indicate that from 57 to 72 per cent of institutionalized mental defectives have speech defects. The most frequently occurring ones appear to involve articulation and voice. There appears to be no evidence to suggest that mentally defective persons are subject to speech defects to which non- mentally defective persons are immune.